question

C T : T D I B T : T E I

B A ( ): K AME ANTHON APPIAH

P : P P

S RL: :// . . / / . 7 9 0.6

JSTOR - - , , , , .

. F JSTOR, @ . .

JSTOR T & C , :// . . /

P P JSTOR , T E I

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

Chapter Three

u The Demands of Identity

LEARNING HOW TO CURSE—THE STRUCTURE OF SOCIAL

IDENTITIES—MILLET MULTICULTURALISM—AUTONOMISM,

PLURALISM, NEUTRALISM—A FIRST AMENDMENT EXAMPLE:

THE ACCOMMODATIONIST PROGRAM—NEUTRALITY

RECONSIDERED—THE LANGUAGE OF RECOGNITION—THE

MEDUSA SYNDROME—LIMITS AND PARAMETERS

LEARNING HOW TO CURSE

In the summer of 1953, a team of researchers assembled two groups of eleven-year-old boys at adjoining but separate campsites in the Sans Bois Mountains, part of Oklahoma’s Robbers Cave State Park. The boys were drawn from the Oklahoma City area and, though previously unacquainted, came from a fairly homogeneous back- ground—they were Protestant, white, middle-class. All this was by care- ful design. The researchers sought to study the formation of in-groups and out-groups—the way that tension developed between them and the way it might be alleviated—and the Robbers Cave experiment has justly become something of a classic in the social sciences.

The camp area was heavily wooded and completely isolated. At first, each group was unaware of the other’s existence. Only after the boys were allowed to settle in on their own for a couple of days did the staff members tell each group that there was another camp of boys nearby. The two groups of boys promptly challenged each other in competitive sports, like baseball and tug-of-war, as well as the less obviously appeal- ing activity of “tent pitching.” Soon—and this was perhaps the study’s most dramatic finding—tempers flared and a violent enmity developed between the two groups, the Rattlers and the Eagles (as they came to dub themselves). Flags weren’t just captured but burned and shredded. Raids were staged on the other group’s cabin; property was disarrayed,

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 63

trophies stolen. Staff members had to intervene when one group of boys prepared themselves for a retaliatory raid by arming themselves with rocks.1

A less dramatic, but, for our purposes, equally intriguing, develop- ment was also recorded. It starts with the self-assigned labels of the two groups: the Rattlers and the Eagles. The groups did not arrive with these names; nor did it occur to group members that they needed a name, until they learned about the presence of another group on the campgrounds.2 Among the Rattlers, an ethic of “toughness” had arisen, after it emerged that one of the higher-status boys in the group had stoically endured a minor injury without telling anyone about it. Cursing, for equally contingent reasons, also became commonplace in this group. When the Eagles won a baseball game against the Rattlers, they came, during a postgame conversation, to attribute the victory to a group prayer they’d offered before the game. After further deliber- ation, the Eagles decided that the Rattlers’ tendency to curse had con- tributed to their defeat as well. “Hey, you guys, let’s not do any more cussing, and I’m serious, too,” one Eagle said to the others, and the proposal won general approval.3 In the course of a subsequent football game, the Rattlers (who won narrowly) engaged in clamorous jeering and boasting. Rather than respond in kind, the Eagles decided that yelling in front of the Rattlers would bring bad luck: they came to refrain not only from cussing but from bragging. These differences were reflected in the way the groups described each other. To the Rat- tlers (in their internal discussions), the Eagles were “sissies,” “cow- ards,” “little babies.” To the Eagles, the Rattlers were a “bunch of cuss- ers,” “poor losers,” and “bums.”4 One group saw itself, and was seen, as prayerful, pious, and clean-living; the other as boisterous, tough, and scrappy.

And all this arose in just four days. The Robbers Cave study, for a student of identity, is a bit like those

origins-of-life experiments where scientists direct a bolt of artificial lightning at a solution meant to represent the earth’s primordial ocean. It’s a long way from those simple amino acids to Noah’s ark. Nor would one want to claim that the nascent trait divergence between the Rattlers and the Eagles does justice to the welter of deep social and linguistic diversity human beings have evolved over the millennia. But as a snap-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

64 u C H A P T E R T H R E E

shot of “ethnogenesis,” there is clarity in its lack of complication, start- ing with the priority of identity to culture. We often treat cultural differ- entia as if they give rise to collective identities; what happened at Robbers Cave suggests we might think of it the other way around.

Comparative scholars of ethnicity have certainly provided no short- age of confirming examples; their reports suggest a similar dynamic of antagonism, lumping even as it splits. The Malay came to know one another as such only after, and in opposition to, the arrival of the Chi- nese; the Hindu became Hindu only when the British created the class in the early nineteenth century, to take in those who weren’t members of the famous monotheisms, and the identity gained salience only in opposition to South Asian Muslims.5 As Jean-Loup Amselle, the French anthropologist, says (in his French way), cultural identities arise, in the first place, from a “structured field of relations,” which is to say that they might be seen, in the first instance, as the consequence, not the cause, of conflicts. “Culture is important in the making of ethnic groups,” Donald Horowitz says, less grandly, “but it is more important for providing post facto content to group identity than it is for provid- ing some ineluctable prerequisite for an identity to come into being.”6

And so I’ll be proceeding in this chapter on the not-uncontroversial assumption that differences of identity are, in various ways, prior to those of culture (a view I’ve argued elsewhere and will defend further in chapter 4).

Among the things we may take from the story of the Robbers Cave experiment is that identity allegiances can be easily conjured into being; and that (if we needed reminding) the Other may not be very other at all. We also know that identity as a social form is no less powerful for all that. Though we may be a society of individuals, in classical liberal terms, the abstraction of that term omits a great deal that matters to us, as individuals and as members of identity groups. Does the liberal goal of equal concern rule out, or require, the acknowledgment of peo- ple as the bearers of identities? If identity may be acknowledged, what sort of political demands can we validly make as members of a collective identity, as opposed to members of a polity? These are the sorts of questions I want to pursue in this chapter. But before exploring what identity demands, I shall say something about what identity means.

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 65

THE STRUCTURE OF SOCIAL IDENTITIES

The contemporary use of “identity” to refer to such features of people as their race, ethnicity, nationality, gender, religion, or sexuality first achieved prominence in the social psychology of the 1950s—particularly in the work of Erik Erikson and Alvin Gouldner. This use of the term reflects the conviction that each person’s identity— in the older sense of who he or she truly is—is deeply inflected by such social features.7 And it is a fact of contemporary life that this conviction is increasingly prevalent. In political and moral thinking nowadays it has become commonplace to suppose that a person’s projects can be expected to be shaped by such features of his or her identity and that this is, if not morally required, then at least morally permissible.

To be sure, not every aspect of the collective dimension of someone’s identity will have the general power of sex or gender, sexuality or na- tionality, ethnicity or religion. What the collective dimensions have in common, as I mentioned in chapter 1, is that they are what Ian Hacking has dubbed kinds of person: men, gays, Americans, Catholics, but also butlers, hairdressers, and philosophers.8

Hacking relies on a crucial insight about “kinds of person,” which is that they are brought into being by the creation of labels for them. So he defends what he calls a “dynamic nominalism,” arguing that “nu- merous kinds of human beings and human acts come into being hand in hand with our invention of the categories labeling them.”9 (It is not incidental that the Rattlers and the Eagles came into being along with their designations.)

Hacking begins from a philosophical truism that finds its most in- fluential formulation in Elizabeth Anscombe’s work on intention: ac- tions are intentional “under descriptions”; in other words, action is conceptually shaped.10 What I do intentionally is dependent on what I think I am doing. To use a simple example, I have to have a wide range of concepts for my writing my name in a certain way to count specifically as “signing a contract.” It follows that what I can do inten- tionally depends on what concepts I have available to me; and among the concepts that may shape my action is the concept of a certain kind of person and the behavior appropriate to a person of that kind.

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

66 u C H A P T E R T H R E E

Hacking himself offers as an example Sartre’s brilliant evocation in Being and Nothingness of the Parisian garçon de café, with his studied air of alertness and solicitude.11 Our own Mr. Stevens, for his part, was driven, in thinking about whether he should develop his bantering skills, by the thought that he is a butler and that banter is a butler’s sort of skill.

The idea of the butler lacks the sort of theoretical commitments that are trailed by many of our social identities: black and white, gay and straight, man and woman. So it makes no sense to ask of someone who is employed as a butler whether that is what he really is. Because we have expectations of the butler, it is a recognizable identity. Those ex- pectations are, however, about the performance of the role; they depend on our assumption of intentional conformity to the expectations.

But with other identities—and here the familiar collectives of race, ethnicity, gender, and the rest come back into view—the expecta- tions we have are not based simply on the idea that those who have these identities are playing out a role. Rightly or wrongly, we do not think of the expectations we have of men or of women as being simply the result of the fact that there are conventions about how men and women behave.

Once labels are applied to people, ideas about people who fit the label come to have social and psychological effects. In particular, these ideas shape the ways people conceive of themselves and their projects. So the labels operate to mold what we may call identification, the pro- cess through which individuals shape their projects—including their plans for their own lives and their conceptions of the good life—by reference to available labels, available identities. In identification, I shape my life by the thought that something is an appropriate aim or an appropriate way of acting for an American, a black man, a philoso- pher. It seems right to call this “identification” because the label plays a role in shaping the way the agent makes decisions about how to con- duct a life, in the process of the construction of one’s identity.

We can describe the relation between identification and identity with a little more precision. In particular, every collective identity seems to have the following sort of structure.12

First, it requires the availability of terms in public discourse that are used to pick out the bearers of the identity by way of criteria of ascrip-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 67

tion, so that some people are recognized as members of the group— women, men; blacks, whites; straights, gays. The availability of these terms in public discourse requires both that it be mutually known among most members of the society that the labels exist and that there be some degree of consensus on how to identify those to whom they should be applied. Let us call a typical label for a group “L.”13 This consensus is usually organized around a set of stereotypes (which may be true or false) concerning Ls, beliefs about what typical Ls are like, how they behave, how they may be detected. Some elements of a stereo- type are normatively derived: they are views about how Ls will probably behave, rooted in their conformity to norms about how they should behave. We can say, in a convenient shorthand, that there must first be a social conception of Ls. Stereotypes are rough-and-ready things, and there may be different conceptions of Ls associated with different indi- viduals or groups within the society. For a social conception to exist, it is enough that there be a rough overlap in the classes picked out by the term “L,” so there need be no precisely agreed boundaries, no determi- nate extension; nor is it necessary that the stereotypes or criteria of ascription be identical for all users of the term. We need not worry that the exact boundary between women and men is not agreed upon (do F-to-M transgendered folk count as men all along, or only after surgery, or never?), or that, even given a full specification of his affectional life and sexual habits, it might well not be universally agreed whether or not Shakespeare was what we now call “straight.” One cannot, there- fore, always speak of the content of a social conception: sketching a social conception requires an ethnography of ways of conceiving of Ls, one that recognizes especially that different stereotypes of Ls may tend to be held by people with different social positions. African Americans, for example, may well have characteristically different social concep- tions of a black identity from others in the United States; and homosex- uals may tend to conceive gay identity differently from heterosexuals. Now, many people have the idea that the normative content of an iden- tity should be determined essentially by its bearers. Even if that is true— which I doubt, since recognition by people of other identities is often a proper source of their meaning—this would still mean that some people would have the content of their identities determined in part by others; namely, those of the same identity.

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

68 u C H A P T E R T H R E E

A second element of a social identity is the internalization of those labels as parts of the individual identities of at least some of those who bear the label. If the label in question is, once more, “L,” we can call this identification as an L. Identification as an L, as I’ve suggested, means thinking of yourself as an L in ways that make a difference: perhaps thinking of yourself as an L shapes your feelings (so that you respond with pride as an L when an L triumphs); perhaps it shapes your actions, so that you sometimes do something as an L (offering a helping hand to another L, perhaps, who is otherwise a stranger; or restraining your public conduct by the thought that misbehavior will reflect badly on Ls). Often, then, being an L carries ethical and moral weight: the no- tion, say, that Jews ought to help other Jews and should avoid behaving in ways that discredit the Jewish community. And often, too, there are behavioral norms associated with identities that it seems wrong to dig- nify with the epithets “ethical” or “moral”: men (sometimes we say real men) walk this way, hold their hands that way, don’t cover their mouths when they laugh. Eagles refrain from cursing and braggadocio.

Identification, in ways we touched upon in chapter 1, typically has a strong narrative dimension. By way of my identity I fit my life story into certain patterns—confirmation at puberty for a religious identity, tenure in your mid-thirties for a professorial one—and I also fit that story into larger stories; for example, of a people, a religious tradition, or a race. Nor is this narrative element simply a feature of Western modernity. Around the world, it matters to people that they can tell a story of their lives that meshes with larger narratives. This may involve rites of passage into womanhood and manhood; or a sense of national identity that fits one’s life into a larger saga.14 Such collective identifica- tions can also confer significance upon very individual achievements: by way of them, you can think of yourself as the first person of African descent to gain a Harvard doctorate in history, or the first Jewish presi- dent of the United States.

The final element of a social identity is the existence of patterns of behavior toward Ls, such that Ls are sometimes treated as Ls. To treat someone as an L is to do something to her in part, at least, because she is an L (where “because she is an L” figures in the agent’s internal specification of her reasons for the act).15 In the current landscape of identity, the treatment-as that is often in focus is invidious discrimina-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 69

tion: gender, sexuality, and racial and ethnic identity have all been pro- foundly shaped (even, in a sense, produced) by histories of sexism, homophobia, racism, and ethnic hatred. But it is as well to recall that not all treatment-as is negative or morally troublesome: sexuality re- quires responding to people as women and as men, and this means that there are patterns of action toward men and toward women that are constitutive of the standard range of sexual orientations.16 Many benev- olent forms of “treatment-as” are meant to counter malevolent forms of “treatment-as.” (Consider the person who, in the late 1930s, urged her German Jewish friends to leave the Third Reich.) Indeed, that iden- tity-based responses can be morally positive should be uncontroversial: many of the world’s acts of supererogatory benevolence involve treating people as fellow Ls—generosity, then, is often a form of treatment-as.

Where a classification of people as Ls is associated with a social con- ception of Ls, some people identify as Ls, and people are sometimes treated as Ls, we have a paradigm of a social identity that matters for ethical and political life. That it matters for ethical life—in the sense I have stipulated—flows from the fact that it figures in identification, in people’s shaping and evaluation of their own lives; that it might matter for politics flows from the fact that it figures in treatment by others, and that how others treat one will help determine one’s success and failure in living one’s life.

In the case of the butler, conventions of behavior associated with a role are explicitly central: the ascriptions are based on the simple idea that someone who works in grand houses of a certain sort will conform to certain expectations; the expectations are based on the conventions that govern the role of the butler; because of those conventions, acting as a butler means constructing a particular performance; and how you are treated may depend on how well you perform the role (even if there are aspects of butlering that are likely to be appreciated only by your fellow butlers). But for some other identities—as a gay man, for exam- ple—there is more than convention.

For being a gay man is, in part, a matter of having certain desires, and those desires are not something that the gay man has himself cho- sen. You can choose whether or not to play a certain conventional role, and, if all there is to an identity is a conventional set of behaviors, and you are capable of them, then you can chose whether to adopt the iden-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

70 u C H A P T E R T H R E E

tity. But when the criteria for ascribing a certain identity include things over which you have no control—as is the case with gender, race, and sexual orientation—then whether you identify with that identity, whether, for example, you think of yourself as gay and act sometimes as a gay person, is not only up to you.17 As we saw in chapter 1, while someone who has a gay identity is doing more than simply acknowledg- ing the fact that he has homosexual desires, and someone who has an identity as a black person, identifying with his or her African American identity, is doing more than simply acknowledging an African ancestry, it is nevertheless true that they are responding to a fact (about desire or ancestry) that is independent of their choices, a fact that comes, so to speak, from outside the self. Even Sartre’s garçon de café takes up an identity that has a function outside himself: he is taking up a profession that provides a service; he is finding, as Mr. Stevens did in butlering, a way of making a life. (Moreover, the profession that he is taking up, with its intricate conventions and protocols, is not one of his own devising.)

For a long time—since the Enlightenment, we might say—the great liberal struggle was to get the state to treat its members as individuals only, without favoring or disfavoring particular ethnic or religious or gender identities. And many people continue to argue that state ac- knowledgment of such identities is intrinsically illiberal: precisely be- cause the shaping of my life is up to me, the government should seek to constrain my acts independent of my identities. Otherwise the state will be in the business of advantaging and disadvantaging particular identities in ways that encroach upon the individual’s freedom to shape his or her own life. Such skepticism draws some of its appeal from the historical arguments for tolerance; and from the apparent clash be- tween the constraining nature of identities and the liberal ideal of the self-directed individual, of the autarky of the soul.

Others, including many so-called multiculturalists, have argued, to the contrary, that the state must recognize these identities because without them individuals will lack what they need for making a life. To the extent that social identities allow people options for making their lives, these theorists argue, they are a positive part of that process. And their recogni- tion by the state is part of what makes them available for this purpose.

As we’ll see, those who think that the political acknowledgment of identity groups is important have produced various arguments for how

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 71

it should matter. Some focus on the state provision of goods and bene- fits; some focus on the suspension of certain rules or obligations that unduly burden members of certain identity groups. Some people go further and—particularly when these identity groups are associated with “societal cultures”18—embrace the principle of cultural sover- eignty. Multiculturalism, perhaps appropriately, comes in many hues. In the past decade or so, the slogans and rallying cries have multiplied: terms like “differential citizenship” (Iris M. Young); the “politics of recognition” (Charles Taylor); and modus vivendi (John Gray) are each associated with distinct approaches toward the matter.

And there are many who occupy an intermediate position. They think that the sorts of social identities I have mentioned are, indeed, ethically central to our lives, but also that this is an argument for tolera- tion of identities, not for their recognition. That is, they are inclined to a view about these social identities that is analogous to the position on religious toleration popularly associated with the American Founders: so far as is possible, no establishment of identities, on the one hand; but, on the other, free exercise (subject to the constraints of duty and harm) as well.19 This position, too, courts its share of perplexities. What would it mean to allow (the analogue of ) free exercise of them while avoiding (the analogue of ) establishment? The answer, as we’ll see, is far from obvious.

In the next section, I’ll be discussing a relatively uncompromised approach toward the acknowledgment of identities—an approach that sees the state as a federation of identity groups, each to be ceded a high degree of autonomy. Group autonomy, as an ideal, comes in many forms, and I cannot do justice to them all; I have largely confined my attention to those forms that arise in the context of liberal political theory (as opposed to, say, Stalinist ethnic engineering). Still, I hope at least to give a flavor both of what its proponents find appealing in this approach and of what others of us find worrying.

MILLET MULTICULTURALISM

Before proceeding with those questions, however, I think it will be helpful to make two distinctions: one about individual-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

72 u C H A P T E R T H R E E

ism and one about group rights. The distinction about individualism in the sphere of rights is between what we can call ethical individualism, on the one hand, and substantive individualism, on the other. Ethical individualism about rights is a liberal assumption I identified in the preface. It is the view that we should defend rights by showing what they do for individuals—social individuals, to be sure, living in families and communities, usually, but still individuals. Substantive individual- ism about rights is the view that rights must always attach to individu- als: that human rights, as framed in our conventions and in law, should always be the rights of persons, not of groups.

The second useful distinction is between two ways of thinking of group rights. One way sees them as exercised collectively: for this to work in practice, there have to be mechanisms by which the groups can be legally identified and institutions through which their interests can be asserted. If an American Indian tribe has the collective right to run a gaming casino, it must be decided both who belongs to that tribe, and how they should decide whether to exercise it. The right of self- determination is a group right of this sort: and it raises both kinds of questions. Who is a Palestinian, a Kurd, a Tibetan? And how should they decide to exercise their rights? Call group rights of this sort collec- tive rights. A second conception of group rights is the idea that the law, whether national or international, might treat each member of certain groups as being individually entitled to certain claims qua member of the group. For instance, members of the English hereditary peerage used each to be able to exercise their right to a trial by the House of Lords. Call group rights of this sort membership rights.

Membership rights are individual rights in a certain sense: they be- long to individuals. But those who say they are skeptical of group rights often mean to be challenging membership rights. What they are ob- jecting to is the idea that a state should relate to any citizen in virtue of his or her membership of a group rather than simply as a citizen. It was an objection to the membership rights of whites (and the member- ship burdens of blacks) that underlay much of the opposition to Ameri- can Jim Crow and to apartheid. Among the membership rights that have a large body of support are the membership rights of citizens of democratic states: it is widely thought to be fine to treat citizens and noncitizens differently before the law; for example in deciding who may

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 73

take jobs where. Collective rights tend to have more friends, however. Most people think that it is just fine that Utah or the city of Cambridge or the Catholic Church can exercise rights, through the ballot box or (in the case of churches) through whatever consensual internal mecha- nisms they agree upon.

It should be admitted at once that ethical individualism is not, on its face, quite so strong a constraint as one might suppose. If our selves are embedded in social forms—the most commonplace of communi- tarian commonplaces—it might be impossible to treat individuals with equal respect without somehow coming to terms with those social forms. That’s a thought that can be given various formulations, some more exigent than others. It might seem that human flourishing—our individual well-being—demands the flourishing of the identity groups within which the meaning of our lives takes its shape. Perhaps, then, a state cannot treat us with equal respect without striving to respect equally the communities in which we are embedded, and which give our choices their context and content.20 Thus we move swiftly from the equal standing of individuals to the equal standing of identity groups, and, indeed, a homology between identity groups and persons is a sta- ple of certain forms of multiculturalism, those sponsored by what we might dub hard pluralism. (Here, “pluralism” is used in something like Horace Kallen’s sense of the term, which promoted the ideal of multi- ethnic coexistence.) As we shall see, it can be hard to keep clear, in such arguments, whether ethical individualism has been breached.

Hard pluralism is a tendency; I do not use the phrase to delimit a single, coherent doctrine. Many hard pluralists object to the automatic elevation of personal autonomy over group autonomy; some are in- clined to suspect even the attempt to draw a distinction between the two. These pluralists are quick to point out the ascriptive and involun- tary nature of those identity groups with which they are concerned.21

Other hard pluralists find their rationale within the traditional vocabu- lary of liberalism; in particular, they judge the freedom of association to be trumps. And there are those hard pluralists who take liberalism to be discredited by the existence of different ways of life, and, in partic- ular, by staunchly illiberal ones: gazing among the various communities in their midst, these pluralists judge nothing to be trumps, save the need to abjure judgment itself.

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

74 u C H A P T E R T H R E E

To speak of “autonomy” is, of course, to map a term that originally described polities—self-governing city-states like ancient Athens—onto persons. But that Kantian move of treating persons as polities can be reversed; many hard pluralists essentially treat polities as persons. They may suggest that identity groups are entitled to all the rights and protec- tions owed a citizen, and so must be indulged on their own terms, with- out worrying overmuch about how it affects the individuals involved. You can get here by various routes. You can, as David Ingram does, make the case for group rights; you can, like Chandran Kukathas, mobilize the freedom of association; you can, with John Gray, appeal to an overriding principle of modus vivendi, thereby raising a strategy of coexistence to an ideal.22 None of these three consistently respects ethical individualism. Drawing upon these approaches, sometimes in combination, theorists have sought to honor the sovereignty of the group, and to minimize outside interference with its affairs, in a way that has sometimes called to mind the millet system of the Ottoman Empire.23

Thus for hard pluralists, personal autonomy is a good the way Mar- mite on toast is: perfectly fine if you happen to have a taste for it, but no more than a local preference aggrandized by the legacy of empire. Toleration—conceived as a relation between the state and its constit- uent groups—is what matters. They are especially concerned that the state not interfere with groups that don’t accept the value of critical reflection and so force the minority group to become more liberal, a concern that is, as we saw in the previous chapter, widely shared by critics of autonomism. William Galston, urging a policy of “maximum feasible accommodation,” recoils at a suggestion made by Will Kym- licka that a state may have the task of liberalizing certain illiberal com- munities while helping to sustain them. “What Kymlicka calls liberal- ization,” he writes, “will in many cases amount to a forced shift of basic group identity; it turns out to be the cultural equivalent of the Vietnam- era principle of destroying the village in order to save it.”24

Hard pluralism would avoid such enormities by leaving groups free to do just about anything to their members short of physical coercion.25

Kukathas argues, for instance, that parents should be able to do what they want with their kids’ education, so long as it has cultural warrant: “because gypsy custom does not value school, the parents believing that they can educate a child satisfactorily through informal instruction in

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 75

the ways of their culture, only a minority of children receive any formal primary education. Their freedom to associate and live by their own ways, however, would, by my argument, make this permissible.”26 And, as we’ve seen, John Gray has made similar arguments for insulating cultural communities from the encroachments of autonomism. Like the millet system of yore, their approach respects the sovereignty of the cultural constituents, imposing external constraints of orderliness but few internal restrictions on how the members of these communities are to be treated.

Yet this elaborate vision is perched upon a very slender plinth. For the proponents of millet multiculturalism have it as an object of faith that personal autonomy is a (usually “Western”) parochialism while group autonomy (often as a “non-Western” demand) is sacrosanct. Why the latter is any less provincial than the former is never explained. And the matter of “cultural warrant” becomes hugely important, requiring a close scrutiny of the habits of strangers. Suppose that, as members of certain religious or ethnic communities sometimes do, you wish to withhold blood transfusions for your injured son, or subject your daughter to female genital mutilation. The main question, for Kuka- thas, is whether your behavior has the sanction of your group—whether it is licensed by the freedom of association.

Now, one kind of situation where the practices of a group are likely to attract interference (as with the examples I’ve cited) has to do with the young, who do not yet enjoy full autonomy. Kukathas, it seems, doesn’t mind a state that imposes compulsory education upon its Scots- Irish families; it’s the Gypsies (or, as they might prefer to call them- selves, the Roma) who may be exempt. But compulsory education arises as an issue of interference only when someone wants to resist it. And who is to tell the Scots-Irish family that doesn’t see the point to school- ing its children that education is part of its culture—that it is, for them, a value? The fact that the family doesn’t want to school its children, you might suppose, is sufficient proof to the contrary.

A second type of situation where the state is likely to interfere with the affairs of a group arises when a member of that group feels that his or her autonomy is under attack. (The doctrine would have no teeth if it only discouraged the state from doing what ordinary scruples about individual autonomy would discourage the state from doing anyway.)

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

76 u C H A P T E R T H R E E

Suppose a Bengali family in England wants to compel its grown daugh- ter to enter into an arranged marriage. If she declines, she has, eo ipso, rejected her parent’s values, at least in some measure. The state cannot immediately decline to uphold her rights on the principle of her par- ents’ cultural autonomy, for what about her culture? Or, if you doubt (reasonably enough) that an individual, by herself, can have a distinct culture, the culture of her cohort? For upon investigation, you may learn that her attitudes are shared by various young women in the com- munity. How, in short, are we to establish the boundaries of the group deserving deference? One imagines a vast brigade of state-employed ethnographers, tasked with certifying this or that practice as legitimized by this or that social group.

And, practically speaking, even if you wanted to leave members of some group to themselves, to preserve their autochthonous form of life, you’d find it was too late: the nature and shape of the leadership struc- ture of a substate polity—a Pueblo tribal council, in the United States, to take an example I’ll be returning to—will have been profoundly shaped by the policies of the state in question. When leadership is con- tested within the tribe, the state will recognize some bids for authority and not others. In this and other respects, the state can’t not take sides.

Even aside from such concerns, you can be forgiven if you do not find the social picture drawn by the strong pluralists to be terribly ap- pealing. A free assembly of small sovereign entities doesn’t resemble a liberal society; it’s a society modeled on the UN Security Council— its plebiscites open to despot and democrat alike. To explain why the approach has the appeal it does among liberals, one must mention a crucial escape hatch it provides: the right of exit. For those who seek to reconcile group and individual autonomy—who seek to exalt the freedom of association without utterly scanting conventional autono- mist considerations—the right of exit has become a veritable work- horse. As long as a group permits members to leave, a great deal is permitted: if you haven’t exercised that right, you have, in some sense, consented to whatever is likely to befall you. Exit thus promises to dis- solve any number of difficulties. Rather than fussing about how inter- nally democratic a cultural community should be—about balancing this interest against that interest, that right against this right—why not simply require the right of exit, and have done with it? Do what you

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 77

like to your own members, we can then say, so long as you let them leave if they want to. That may appear to comport with ethical individualism, because it means that group choices that are bad for the individual member cannot be imposed without her consent. Unfortunately, this right is not so straightforward an affair.

Is Exit Enough?

For one thing, the right of exit is actually rejected by some hard pluralists. It may sound like a pretty minimal condition, but those who have no time for autonomism regard even this right as too exacting. Gray, for example, sees exit as yet another corrosive liberal imposition. How can you eliminate any form of life that can’t withstand the exercise of free choice by its members, he asks, and still call yourself a pluralist? What if the exit of a few would ruin a community for the majority who endorse it? Then wouldn’t the infringement on individual autonomy be justified, as in the case of the fire-shouting theatergoer? Nor are these merely abstract concerns. It should be admitted that there are social formations that couldn’t survive an opt-out clause; history has seen one or another leisure class that depended upon the services of a class of serfs or slaves: once the slaves opt out, the aristocratic charms of Southern plantation culture are imperiled. Those of us who love the literature of Augustan Rome must recognize its dependence on the re- sources of an involuntary empire. Or consider Athenian democracy: the golden age of Pericles, and the role of involuntary servitude in sus- taining it.27 The right of exit is certainly inconsistent with radical toler- ance of groups; and how you feel about that depends on what sort of priority you give that form of tolerance.

But a more widely shared concern is that, conversely, the formal right of exit is too weak to do the work assigned it. Indeed, once you start thinking about such mechanisms of escape, you can start to wonder whether the “freedom of association” model doesn’t lump together some terribly disparate phenomena. Can a nationality fairly be seen as analogous to a private club? Is an identity group something you can simply resign from?

Consider a situation where there’s a dramatic collision between group and personal autonomy. The Pueblo tribal council doesn’t per-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

78 u C H A P T E R T H R E E

mit religious freedom among members of the Pueblo tribe. It also has control over communal property, which is, for the Pueblo, essentially all property. Now, to take an example that has occasioned litigation: does the Pueblo who is expelled for converting to Protestantism fully enjoy the privilege of exit, when exit entails the loss of all his earthly goods? You may wish to construe the Pueblo community as analogous to any old voluntary community; but it plainly has power over its mem- bers that, say, the Knickerbocker Club or the Park Slope Food Coop does not.28

And barriers to exit needn’t take the form of property deprivation to be formidable. The philosopher and legal scholar Leslie Green points out, “It is risky, wrenching, and disorienting to have to tear oneself from one’s religion or culture; the fact that it is possible to do so does not suffice to show that those who do not manage to achieve the task have stayed voluntarily, at least not in any sense strong enough to undercut any rights they might otherwise have.”29 And, as he sees, the identity groups we’re most likely to take seriously as “experiments in living” are ones in which belonging is an ascriptive, organic affair, the furthest thing from the voluntarist model. “Everything about a culture is an exit barrier,” Jacob T. Levy persuasively observes. “To have a cul- ture where exit is entirely costless . . . is to have no culture at all.”30

Indeed, you may ask what sense it really makes to say you can exit an identity group. As ex-Mormons like to point out, being an ex-Mor- mon has itself become a kind of ethnicity; and something like that condition seems to hold for other ascriptive groups. More to the point, if the unencumbered self is a myth, how can you extricate yourself from the context that confers meaning? After all, it would make little sense to speak of “exiting” your language, especially when it is the only one you have.

But if the mere existence of an exit isn’t enough to justify a policy of millet-like laissez-faire, we’re left without this shortcut and must take the long way around. In particular, we have to take seriously an ap- proach we might call liberal multiculturalism, or soft pluralism. Here the aim is to balance external rights and internal constraints. If strong autonomists are apt to suspect substate groups of restricting freedom, and hard pluralists, contrarily, are apt to celebrate them as sites where state authority is kept at bay, soft pluralists try to salvage something of

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 79

both group and personal autonomy. To what extent the approach suc- ceeds is something I want to consider now.

AUTONOMISM, PLURALISM, NEUTRALISM

In soft, or liberal, pluralism, the individual remains both the terminus a quo and the terminus ad quem: its concern for identity groups is not only motivated by but ultimately subordinated to the well-being of the individual and the bundle of rights and protec- tions that traditional liberalism would accord her. Ethical individualism is meant to have its way. In contrast to the kind of millet multicultur- alism we’ve considered—where state power is made to recede as far as possible, and substate autocracy is indulged, at least short of outright coercion—soft pluralists try to find a point of equilibrium between the rights of individuals and the integrity of intermediate associations. For many of them, moreover, that ideal of balancing has a name: neutrality.

People who share this goal can differ widely in how willing they are to accommodate identity groups. Will Kymlicka, an exemplary soft pluralist, is the most eloquent of those who would justify group auton- omy as a means to the end of individual liberty. He is one of those who doubt that the right of exit is sufficient protection for the individuals involved. Accordingly, he is concerned to strike a balance between two desiderata. On the one hand, he wants to ensure external rights (the rights of the group against the state, and against the incursion of outsid- ers; but also the rights of the group to procure, from the state, certain group-specific benefits). On the other hand, he wants to limit the inter- nal restrictions a group may impose upon the autonomy of the group’s members. In balancing these things, though, he would deny that he is trading off individual interests against collective ones. The flourishing of individuals, in his view, requires the stability and security of their social forms.

And yet the contrast between internal restrictions and external pro- tections quickly breaks down. A group wishes to educate its members only in the minority language: is this external protection from the lin- guistic hegemony of the overlords? Or is it internal restriction, inas- much as it constrains the opportunities of the younger generations,

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

80 u C H A P T E R T H R E E

impedes their right of exit, and so infringes on the autonomy of their future selves? Kymlicka regards the Wisconsin v. Yoder case—in which the Old Order Amish asked for (and received) an exemption from the state of Wisconsin’s requirement that they send their children to school until the age of sixteen—as involving an instance of internal restriction; and yet it could as easily be presented, and doubtless strikes the Old Order Amish themselves, as a matter of external protection of their agrarian way of life. In another familiar type of case, an aboriginal group imposes restrictions on who may buy members’ land: again, this is a safeguard against the encroachments of the outside world that con- strains the autonomy of the group’s members. As Kymlicka admits, the distinction “is not always easy to draw.”31

I mentioned that the American conception of religious tolerance pro- vided one obvious model for how the state should respond to identity groups. Indeed, as the Yoder case should remind us, the project of rec- onciling the two desiderata of our associational affairs—freedom of and freedom from—is most prominently represented by the tradition of First Amendment jurisprudence, by which the U.S. Supreme Court has tried to work out a systematic approach to religious freedom. The chal- lenge starts with the very first sentence of the Bill of Rights: “Congress shall pass no law respecting an establishment of religion or prohibiting the free exercise thereof.” Constitutional interpretation bifurcates along that “or.” In modern jurisprudence, the “free exercise” clause has been taken to require the state to extend special deference to religionists; the “establishment” clause has been taken to forbid it. Now, our moral modernity naturally takes the prohibition of “establishment” as an ex- pression of a larger ideal, namely, that the state should be neutral— neutral among competing conceptions of the good, neutral among the competing interests of identity groups. It is the ideal at the heart of Rawls’s famous, and famously vexed, distinction between “political lib- eralism” and “comprehensive liberalism.”

Famously vexed: skeptics have argued at length that such neutral- ity is incoherent or undesirable or, in a manner of speaking, both— that such neutrality is bad because it favors neutralists over others and so isn’t really neutral at all. Thus, in the context of the First Amend- ment, religious advocates accuse neutralists of elevating a controversial conception to the level of procedure; and they feel oppressed by neu-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 81

tralist strictures that (as they interpret them) exclude religion from the public square.

In one widely discussed version of the argument, Michael Perry has taken aim at the suggestion advanced by theorists such as Thomas Nagel, Robert Audi, and Bruce Ackerman that ideal political discourse should appeal, as much as possible, to shared and publicly accessible forms of reason. To the extent that expressly religious appeals—appeals, for in- stance, to the “revealed truth” of God—shut out and do not seek to engage nonbelievers, they are not ideal specimens of political delibera- tion, of what John Rawls calls “public reason.” What Michael Perry maintains is that to ask one to check one’s religious convictions at the town-hall door is to ask one to “bracket—to annihilate—essential as- pects of one’s very self.” Conversational strictures that purport to be neutral are really anything but: they condemn one’s spiritual self to be “marginalized or privatized,” leaving the devout unable to participate in the political realm on an equal footing with their secular compatriots. Perry, accordingly, seeks to establish that a duly pluralistic conception of the public realm should leave space for “religious-moral discourse.”32

In truth, the adherents of “political liberalism” have never been quite so coercively monistic as their critics often represent them. Certainly the right of the zealous to say what they like is not disputed by neu- tralists; at issue is merely what ideal political discourse ought to sound like. As Rawls insists, public reason, far from requiring citizens to sus- pend or withdraw their religious convictions, in fact presupposes that citizens have a variety of comprehensive conceptions, including reli- gious ones, whose “overlapping consensus” is consistent with the insti- tutions of liberal democracy.33 Indeed, those Rawlsian strictures about the ideal of public reason are perhaps best interpreted as debating tips: as rhetorical advice about how best, within a plural polity, to win adher- ents and influence policies. There’s nothing coercive about such coun- sel. Sectarians may speak however they prefer, but if they seek to win over those who do not already share their sectarian convictions, they will be well advised to appeal, as much as possible, to those norms and premises that are most generally accepted. So the spirit behind these liberal strictures is less Madalyn Murray O’Hair than Dale Carnegie.

If the neutralist ethic is less invidious than some religionists would claim, it does face some rather daunting problems all the same. Thus

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

82 u C H A P T E R T H R E E

Will Kymlicka objects to the notion that the diversity of social forms ought simply to be treated with “benign neglect” by an assiduously impartial state. “In the areas of official languages, political boundaries, and the division of powers, there is no way to avoid supporting this or that societal culture, or deciding which groups will form a majority in political units that control culture-affecting decisions regarding lan- guage, education, and immigration,” he points out. As he elsewhere elaborates: “It is quite possible for a state not to have an established church. But the state cannot help but give at least partial establishment to a culture when it decides which language is to be used in public schooling, or in the provision of state services. The state can (and should) replace religious oaths in court with secular oaths, but it cannot replace the use of English in courts with no language.”34

And, of course, the government—at least in anything resembling an actually existing liberal democracy—is not neutral, cannot be neutral, over a vast range of things. A government science agency bestows grants to avenues of research it thinks are promising and not to others; the language of governance is not Esperanto; decisions are made, and rules are interpreted, in ways that favor some interests and disfavor others: competing rights and interests are balanced, and the balance typically comes out in the favor of one party over another. In such cases, facts and values come together in a theory-saturated mélange, so that one cannot say that a finding of fact was not also a finding of value.

But perhaps this sort of neutrality isn’t the sort that neutralism is really concerned with. In an influential line of argument, Raz and (in more detail) Kymlicka have proposed that the ideal of “neutrality” ap- plies not to effects but to justifications. State actions can never achieve anything like neutrality of consequences; but states can strive toward neutrality of rationale. That is, liberal neutralists realize that the state will act in ways that have nonneutral effects, but want to be assured that such acts have a motivation other than (say) reshaping our religious identities. And there is a range of examples where the neutrality-of- justification test can provide some conceptual purchase. A law regulating the slaughtering of animals will not pass muster if it clearly has no real purpose except to target practitioners of Santeria, as the Supreme Court found of certain ordinances passed by the city of Hialeah, Florida.35 A case like that seems straightforward. But, as we’ll see, most cases aren’t.

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 83

What First Amendment jurisprudence offers isn’t so much wisdom as an illustration of the perplexities that arise from all efforts to honor both state neutrality and group autonomy. At the same time, religious groups are among the more salient buttresses of identity in the West; this isn’t merely an analogy but an example—indeed, for many, a paradigm.

A FIRST AMENDMENT EXAMPLE: THE ACCOMMODATIONIST PROGRAM

A little background may be helpful. Over the past cou- ple of decades, a diverse and influential group of so-called accommoda- tionist scholars (among them Michael McConnell, Frederick Gedicks, W. Cole Durham, Harold Berman, David Smolin, and Stephen L. Car- ter) have sought to strengthen the free exercise clause; many have, ac- cordingly, sought to reverse what they see as a grave misinterpretation of the establishment clause that took root in the Warren Court years. It is a cause that has enjoyed impressive, though partial, success.

As the accommodationists like to point out, the notion of a strict separation between church and government is a recent one. Although Jefferson spoke of a “wall of separation between church and state” in his 1802 letter to the Danbury Baptists, the First Amendment, speaking only of what laws Congress could pass, left the states to determine their own church-state relations. Only in 1947, with Everson v. Board of Edu- cation, did the Supreme Court pronounce nonestablishment to be bind- ing upon the states as well as the federal government. And it is with Everson that modern “separationist” jurisprudence originates. In words quoted as often with opprobrium as approval, Justice Hugo Black’s opinion declared not only that the First Amendment “has erected a wall between church and state,” but also that this wall “must be kept high and impregnable. We could not approve the slightest breach.” In fact, the Court has long since retired Justice Black’s rhetoric of separation. Since 1971, Black’s high and impregnable wall between church and state has been replaced with what the Burger Court called a “blurred, indis- tinct, and variable barrier depending on all the circumstances of a par- ticular relationship.” In a doctrine first introduced in Lemon v. Kurtz- man in that year, and still apparently in effect, disputed legislation

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

84 u C H A P T E R T H R E E

would be submitted to a three-pronged test. First, the legislation must have a secular purpose; second, its primary effect must neither promote nor inhibit religion; and third, it must not foster excessive entanglement of government with religion. Violation of any of these conditions would render a statute unconstitutional under the establishment clause.

The Lemon test seems an earnest effort to achieve justificatory neu- trality, but, across a surprising swath of the political spectrum, its de- fects are widely agreed upon and widely bemoaned. Consider the “secu- lar purpose” proviso. As subsequent decisions have interpreted it, that “purpose” refers to the subjective motivation of the legislators. But if religious scruple motivates a legislator to introduce funding to shelter the homeless or feed the hungry, should the resultant statute really be declared unconstitutional? Next, consider the “neither promote nor inhibit” proviso. If a government refuses to provide the sort of assis- tance to religious schools that it provides to other private schools, is it not inhibiting religion? But if it does provide such assistance, is it not promoting religion? Here the state is in a no-win situation.36 Finally, consider the “no excessive entanglement” proviso. To ensure the reli- gious neutrality and secular intent required of the first two prongs, state funding for church-run social services may require that the services be monitored to ensure they do not involve religious inculcation; but such monitoring, even if the church acceded to it, may be forbidden as “ex- cessive entanglement.” Here the church is in a no-win situation. So it’s little wonder that the Lemon doctrine has pleased neither accommoda- tionists nor separationists. Until a majority can agree upon a replace- ment, however, it persists, as Justice Scalia once complained, like “some ghoul in a late-night horror movie” that won’t stay dead.

What Everson did for the establishment clause, Sherbert v. Verner did for the free exercise clause in 1963. It propounded the doctrine that the application of a law that burdened an important religious practice was permissible only if justified by compelling state interest. The offer of religious exemptions from generally applicable laws was not new: much “temperance” legislation, for example, accommodated Catholics and Jews by exempting the sacramental use of wine; selective service legisla- tion made similar allowances for religious objectors. What was new in Sherbert was the notion that the Constitution might require such a reli- gious exemption. In this case, a Seventh-Day Adventist, discharged for

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 85

failing to work on Saturday, his day of Sabbath, sued the state for unem- ployment benefits and won on free exercise grounds. Unfortunately for the vitality of the principle, almost all subsequent applications of the mandatory exemption rule involved dissident Sabbatarians pursuing unemployment compensation cases.

To complicate matters further, the two religion clauses have often been pitted against each other, and without any accepted techniques of mediation. On the one hand, the Court has continued to declare unconstitutional, on free exercise grounds, the denial of state unem- ployment benefits to religionists who lost work because they refused to work on their Sabbath. On the other hand, in Thornton v. Caldor (1985), the Court ruled against a Presbyterian petitioner in declaring unconsti- tutional, on establishment grounds, a Connecticut statute that wrote this evident principle of Sabbatarian accommodation into law.

As it happens, the real cause for accommodationist alarm isn’t estab- lishment clause jurisprudence, where the religionists have generally had their way. The real cause for alarm has been the abandonment of the doctrine of free exercise exemptions—the end of the old Sherbert test requiring exemptions, in the absence of a compelling state interest to the contrary, from laws that interfere with a central aspect of religious practice. And here civil libertarians and religious accommodationists alike stand united in opposition. It might be said that a conservative fusillade cut the establishment clause down to size, and the free exercise clause was an accidental casualty.

The contrast Galston and others assert between autonomism and diversitarian tolerance is perhaps visible here. What many accommoda- tionists argue against is the tradition—John Locke’s Letter Concerning Toleration is its touchstone—that sees the primary value enshrined in the doctrine of religious freedom to be the autonomy of the individual conscience. On the contrary, they argue, religious freedom must further the autonomy not of the individual but of the church, the collective entity, which is much more than the sum of its parts.37 The claims of the religious institution against the state, rather than the claims of the individual religious dissenter as such, must be taken as primary.

Such religious communitarianism is often buttressed by a frankly Madisonian rationale. The state that accommodates religion has ac- cepted an important check on its own power. Rather than being mere

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

86 u C H A P T E R T H R E E

aggregations of individuals, religions should be seen as commun- ities often engaged in “interposing the group judgment against the judgment of a larger society” (in Stephen L. Carter’s words); because they provide a potential alternative source of authority, religions func- tion as a bulwark against state tyranny.38 That, in the relevant sense, is what religion is for: and, consequently, what free exercise accommoda- tions are for.

This communitarian perspective asks us to give the widest berth to the autonomy of religions as self-regulating institutions. (And let me stress that this is not a view shared by all accommodationists: it is one strain of thought among others.) It asks us to accommodate, ceteris paribus, the demands of corporate worship over the demands of the individual religious dissident. It asks us to be wary about interposing secular judgments about racial or sexual equality upon dissenting be- lief communities. The result is a sort of jurisprudence of group auton- omy.39 No doubt one could often arrive at the same place through resolutely individualist considerations (traveling by routes we’ve ex- plored); free exercise for me might require a corporate body with au- thority over its members. But there will surely be cases in which it matters whether one takes the robustness of the religious community to be the prior objective.

Many accommodationists are also concerned that courts often fail to respect religious beliefs—fail to respect what Carter terms the “alterna- tive epistemology” of the church. What we haven’t understood, we’re told, is that religion demands “an epistemology of its own”—that it is “really an alien way of knowing the world—alien, at least, in a political and legal culture in which reason supposedly rules.”40 And yet the at- tempt to protect the church not merely politically but epistemologically may tend to undermine the call for greater inclusion of religion within the public sphere.

Consider the conflict that has sometimes arisen between Jehovah’s Witnesses and the courts on the subject of blood transfusions. Jehovah’s Witnesses believe that such transfusions violate the Divine injunction against ingesting blood and—even if received unknowingly—will for- ever doom their souls to perdition.41 As accommodationists have re- provingly noted, the courts do not take these convictions into account. Although courts generally respect the expressed wishes of an adult to

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 87

refuse a potentially lifesaving transfusion, they do so on general grounds of individual autonomy, irrespective of the religious nature of the request. Consequently, the courts are inclined to make no such allowances for a member of the sect who has arrived at the hospital unconscious, or who has not reached the age of majority.

Now then: can the law, in failing to take religion into account, be accused of discriminating religiously? Apparently so. This is Carter’s analysis: “By forcing the Witness to live and be damned rather than permitting her to die and be saved, the state is necessarily treating her religious claim not as irrelevant, but as false.” And that, he stresses, is a crucial distinction. Liberal epistemology, captive to empiricism, cannot take seriously a very serious truth-claim indeed. Once again, liberal- ism’s vaunted neutrality reveals itself to be anything but. “Liberal neu- trality,” Michael McConnell writes, “is of a very peculiar sort,” for it proceeds “as if agnosticism about the theistic foundations of the uni- verse were common ground among believers and nonbelievers alike.”42

We are thus invited to see things from the religionists’ viewpoint, respect their version of reality, and defer to their “alternative epistemol- ogy”: the temporal authority of the medical practitioner may thus be trumped by the spiritual authority of the sect. The ultimate result of such epistemic forbearance, however, goes beyond protecting the sec- tarian from unwelcome interference; the ultimate result is to erase the legal distinction between spiritual and temporal considerations. If the judge and the medical practitioner can be compelled to act as if the sectarian’s beliefs are true, on what grounds can we deprive belief com- munities of—or, conversely, exempt them from—the agencies of tem- poral adjudication?

To see where this leads, consider another example. In the early eight- ies, as Carter reports, New York State passed “legislation that, in effect, requires an Orthodox Jewish husband seeking a civil divorce to give his wife a get—a religious divorce—without which she cannot remarry under Jewish law.”43 The accommodationists show us how such a stat- ute is to be defended. Requiring the granting of a get in such cases can now be seen as a humane provision, and an instance of the state treating religion seriously, rather than as a hobby, as a matter of conviction, rather than velleity. As such, it exemplifies the sort of jurisprudence of relativism that guided us through the Jehovah’s Witness case.

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

88 u C H A P T E R T H R E E

The trouble is that this jurisprudence of relativism is on a collision course with the jurisprudence of group autonomy. Under the former, what the Lemon test calls “excessive entanglement” is positively re- quired. If the courts have the authority to enjoin gets from reluctant religionists, why stop there? Why not review excommunication deci- sions made by Mormons, or the dreaded commands to shun apostates issued by Jehovah’s Witnesses? Did the preacher wrongly damn your soul to hell everlasting? The state, by this logic, should be able to offer at least the injunctive remedy of preacherly recantation. In short, there’s a clash between the proposed politics of religion and the proposed epis- temology of religion. The ideal of regarding religious truth-claims as self-validating cannot be squared with the ideal of protecting institu- tional autonomy. One directs insulation; the other forbids it.44

Indeed, every state is constantly taking sides on creedal matters, a point Kent Greenawalt makes with marvelous concision: “A court or- ders a state to desegregate its schools, the country goes to war, educa- tional funds are made available equally to men and women. The gov- ernment has implicitly rejected religious notions that (1) God wishes rigid racial separation, (2) all killing in war violates God’s command- ments, (3) all women should occupy themselves with domestic tasks. A vast array of laws and policies similarly imply the incorrectness of particular religious views.”45 And the most vigorous efforts to eliminate such disagreement between church and state seem to disserve both.

NEUTRALITY RECONSIDERED

To make progress here we need, I think, to see what is right and what is wrong about talk of liberal “neutrality.” The ideal, put negatively, is that governmental action, including but by no means limited to legislation, should not exhibit partiality toward some sub- group of the nation; put affirmatively, then, states should be neutral among identities. We have already seen that state action cannot be neu- tral in its effects; necessarily, many state acts will have differential im- pacts on people of different identities, including religious identities. Once we turn to the other obvious alternative—neutrality of aim or justification—however, we face immediately the question of how to

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 89

identify a state’s aims or justifications. What is it for a state act to be undertaken for this or that reason or to be guided by this or that intent?

American courts regularly appeal to the intent of legislators in guid- ing statutory interpretation and in scrutinizing statutes to see whether they pass constitutional muster, and there is a tradition of thought as to how to carry out this task. Such notions of legislative intent may be useful legal fictions. But it does not in general make sense to suppose that a legislature has an intent in passing a law. Legislation is a political process, in which deals are cut and compromises made. In both the public and the private deliberations about any statute many inconsis- tent reasons will be offered for framing a clause one way or another; many suggestions, not all of them consonant with one another, will be offered as to what the overall aims of the statute are. To extract from this mishmash of mixed motives a singular coherent intention will usu- ally be impossible.46

You might think that the problem is easier when you are inquiring into just those intentions relevant to establishing neutrality, for these fall into a relatively focused domain. A law should not be aimed to advantage or disadvantage people of a certain identity (narrower than the whole political community) as such: that a legislator is or is not individually motivated in this way might be something we could assess. We could propose, for example, that the presence of such motives in a majority of those enacting the measure established it as nonneutral.

This principle seems, however, to be both too strong in some respects and too weak in others. It is too strong because the mere presence of such a motive does not seem by itself to be enough to discredit an official act. We saw this when we discussed the “secular purpose” pro- viso of the Lemon test. If there is a good impartial, unbiased, reason for a legislative act (or, indeed, for any act of a state official) why should we deny that the act is neutral? If there is a good neutral reason opera- tive, the fact that there is a biased one as well seems to discredit the actor, not the act.

We can see this both in the legislative case and in judicial and execu- tive action. Let’s call a set of reasons for an official act that, taken to- gether, would be strong enough to motivate that act a set of sufficient reasons. It might be that every member of some legislature has a private grudge against a group and knows that a law will tend to affect them

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

90 u C H A P T E R T H R E E

adversely. But if each of them is also motivated by a set of neutral suffi- cient reasons, and the neutral reasons would have led them to vote that way without the extra element of bias, then there seems no reason to impugn the law that they enact. We might think the law was unobjection- able, even if they were not motivated by such a neutral set of sufficient reasons, so long as there were such a set of reasons; that is, so long as there were good enough neutral reasons to motivate the law, even though they did not in fact motivate it. The same applies, mutatis mutandis, to other official acts. If a judge dislikes Catholics and takes pleasure in sentencing a Catholic for a crime, this surely reflects badly on her. But provided there are adequate grounds for the sentence that do not reflect that bias—and especially if those are the grounds she gives for her sen- tence—there seems no reason to claim that the sentence reflects state bias; just as a law that would have been passed if there had been no such biased motives is not vitiated by the presence of subjective bias.47

On the other hand, as I say, the proposal that we count as nonneutral only those acts that are motivated in part by bias seems too weak. For one kind of state nonneutrality consists in passing laws that negatively affect some subcommunity not because legislators intend to do so, but because they have not taken the trouble to examine what the impact of the law will be on them. Sometimes nonneutrality shows up, in other words, in negligence. This is evident in the case of religion. Suppose that, motivated by the desire to protect both husbands and wives from the possibility that their partners would will them too little of the mari- tal property (in order, say, to provide for a secret lover), a state were to require that people must provide for their spouses in a certain way. Suppose that, unbeknownst to the legislators, that rule was in some way inconsistent with the rather specific commands of the Koran. (I should make it plain that the Koran is itself concerned that husbands make adequate provision for their widows.) It would be reasonable, I think, for an American Muslim to object that such a law offended against the idea of neutrality, on the grounds that it would be possible to achieve its stated aim without offending against (at least this aspect of ) sharia. (If, per contra, there were a compelling state interest that did not admit of accommodation, a state could legitimately ignore such an objection.) Something like this notion lies behind the idea, in Ameri- can antidiscrimination law, of “disparate impact.” Where a policy with

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 91

a certain express aim tends to disadvantage a historically disadvantaged racial minority—and thus has a “disparate” racial impact—and an al- ternative policy that would achieve the same aim and would not so disadvantage them is available, the Court has sometimes held that the policy may be legally barred as discriminatory, even without any show- ing of an intent to discriminate.48

You might think that, in speaking of legislators’ aims, we should address ourselves to a bill’s rationale—what it ostensibly aims to do— rather than to the subjective motivations behind its support. But, as the matter of disparate impact makes clear, a law can fail to be neutral even if its stated rationale is scrupulously so. And sometimes public acts that profess neutrality are nonneutral out of intent, not merely inadvertence. A city council might, in promulgating a rule, proclaim a concern about cruelty to animals but really seek to rein in an unpopular form of religious worship. So a law could be facially neutral—neutral in its rationale and its declared justification—but, in both motivation and effect, be discriminatory. Distinguishing between aims and intent, or justification and motivation, doesn’t seem to clear things up.

In the face of these objections, there might still, I think, be a way to salvage something of the idea of neutrality among identities. It is to insist that state acts should treat people of diverse social identities with equal respect. Where an act disadvantages people of identity L, they can reasonably ask whether they could have been treated better and whether they would have been, had they not been regarded as Ls. In all the cases I have just considered, if the answer to that question is yes, we were inclined to treat the case as one of nonneutrality; where it was no, we were not. I shall call the ideal implicit in this test—that state acts shouldn’t disadvantage anyone in virtue of his identity—neutrality as equal respect.

(Given the aims of this book, it is worth underlining here how differ- ently identity works, on this liberal conception, for the state and for the individual. For the individual, as we have seen, that someone is an L can be a perfectly proper reason for treating him differently from others. Because identification constitutively makes being-an-L figure among our reasons for action, neutrality among identities, far from being an attractive moral ideal, is barely intelligible for us as individu- als. That it may be required of the state is a reflection of something

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

92 u C H A P T E R T H R E E

special about public reasons: that they address us all equally as citizens. I have called this notion of neutrality “neutrality as equal respect” in part to indicate how close it is in substance to the ideal of equality that is a long-established part of the practice of liberalism. It is the equality that traveled with liberty and fraternity in the French Revolution; the equality of the American Founding Fathers. I shall have more to say about the ways in which our private reasons are free from the obliga- tions of this sort of neutrality in chapter 6.)

I have expressed skepticism about identifying the aims of state acts: you might, therefore, think the right question to ask is not whether someone would have been treated better if she hadn’t been regarded as an L, but whether she would have been treated better if she hadn’t been an L. So let me spell out why, on our test, the fact that someone is regarded as an L must explain her disadvantage. Consider a particular case where the treatment of a minority disadvantages them in relation to a majority. Left-handed people live in a public world where many things—scissors, door handles, cabinet doors—are configured in a way that suits the right-handed. They are disadvantaged by this fact; and the fact that they are left-handed plays a role in the explanation why. But the fact that they are regarded as left-handed plays no such role, at least in most places in the industrialized world, where there is nowadays little prejudice against the left-handed. The reason they are disadvan- taged is that some things have to be done in either a left-handed or a right-handed way, and right-handed people are in the majority.49 Simi- larly, when it comes to Sabbatarian issues, the fact that our weekend coincides with the religious requirements of the Christian majority doesn’t display a failure of neutrality as equal respect, provided that what accounts for the fact is that it suits a majority (and that, for coor- dination reasons, people cannot be permitted to take their two days in seven on whichever days they choose) and not that some minorities are disadvantaged by it. If a majority of Americans became Muslims and Friday mosque became a majority institution, it would not reflect a lack of regard for Christians if we shifted the days when government offices were closed to Friday and Saturday.

This formulation allows us to see something important in the de- fenses of liberal neutrality that focus on there being neutral reasons for a state policy. Thomas Nagel, in an influential paper titled “Moral Con-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 93

flict and Political Legitimacy,” argued that questions of neutrality arise where the state exercises its coercive power against the will of a citizen. He suggested that in these circumstances, the citizen is entitled to a reason (or a set of reasons) for the act that legitimated it: by which he meant a reason the coerced citizen should accept as a ground for the act.50 (Nagel rightly insisted that we need not be bound by whether the citizen in fact accepts our reason, for otherwise people could, in effect, veto policies simply by being unreasonable; he required only that there be a reason that the coerced citizen should accept, or, equivalently, that she would accept if she were reasonable.)

There are three immediate problems with this proposal as it stands, none of which is a problem for neutrality as equal respect. The first is that citizen and reason-giver may disagree about what is reasonable. In these circumstances—especially, for example, in matters of religious belief—even if the state’s view is in fact reasonable, ignoring the citi- zen’s objections may not adequately display equal respect for all reli- gions, given our ordinary understanding of liberal religious toleration. In the sphere of religion, one of the issues in dispute is often what it is reasonable to think and do. Neutrality as equal respect is not open to this objection. Under neutrality as equal respect, we can make sure that we don’t treat people of one religion (as such) worse than others, even if we believe their views are unreasonable—even if they are, in fact, unreasonable. Many Jehovah’s Witnesses, as I mentioned, believe that having a blood transfusion will lead to their damnation. Even knowing this, we might pass a law that required the provision of blood transfu- sions to unconscious persons who needed them, because a policy re- quiring us to establish consent would endanger the lives of many. Of course, if we thought blood transfusions would in fact lead to damna- tion, we would have a decisive reason not to adopt this policy. But we do not believe this. We mostly do not think it is even reasonable to believe this. On the account of neutrality in terms of reasonable accep- tance, if we are coercing people who are unreasonable, we have no obligation to provide justifications that satisfy them. We must merely offer reasons that they would accept if they were reasonable. Here, so it seems to me, Nagel’s neutrality demands too little. Under neutrality as equal respect, Jehovah’s Witnesses will also sometimes be coerced into doing what we think is best for them, even though, because their

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

94 u C H A P T E R T H R E E

beliefs are unreasonable, they think we are doing them great harm: but we will be thwarting their wills for some good reason, having reflected on whether we could adopt a policy that did not thwart their aims, and so the fact that they are Witnesses does not explain why we went against their aims.51 Neutrality as equal respect doesn’t require us to feign ag- nosticism about the beliefs of our fellow citizens or avoid relying on controversial claims: it merely asks us to avoid offending the beliefs of minorities as much as we can.

A second problem becomes clear when we realize that the existence of a reason that the citizen should accept—a neutral reason—can be construed in two different ways. It might mean that there was a neutral all-things-considered reason: a reason strong enough, in the light of all the countervailing considerations, to justify the policy. But to insist on this would mean that we could never legally proscribe an act where there were citizens who rejected our reasons for proscribing it. For otherwise the law would be coercing some people (whether by punishment or the mere threat of punishment) who did not recognize our grounds for legislating as adequate. Given the fact of pluralism, this would make legislation in many areas impossible. Nagel’s argument construed the existence of a neutral reason for a policy in the other way—as meaning only that there are considerations shared by the state and by the person coerced that favor that policy; but this makes neutrality too easy to achieve. We could, for example, prohibit the wearing of turbans, on the grounds that anti-Sikh sentiment leads to some acts of violence. This is certainly a reason for the policy, and, since Sikhs obviously have reasons to want Sikhs protected from the assaults of bigots, it is a reason a rea- sonable Sikh would accept. But he would go on to say that the wearing of turbans was too important a matter to be banned for this reason. Why, he might ask, shouldn’t we beef up policing of bigots instead?

Now, the reasons Sikhs have for wearing turbans are not reasons that most of the rest of us recognize. And that, of course, is the point. Even if we had sufficient nonreligious reasons to support a policy that coerced members of a religious group, absent consideration of their religious reasons, and even if they took those nonreligious reasons to be valid, they might still believe that these reasons were overridden by religious considerations. Nagel’s proposal would count us as neutral because we had sufficient reasons, shared with the Sikhs, for the policy: but that is

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 95

only because we do not share the reasons that they would take to be dispositive. If the Sikh asks whether his religious duty (to wear a turban) was ignored because he belonged to a religious group with which others have little sympathy, I think a fair answer might sometimes be yes. Neutrality as equal respect might then lead us to take the fact that the wearing of turbans is central to Sikh life as grounds for carving out an exception for them, as we would carve out exceptions, where practica- ble, for other religious practices. Sikhs, the principle is, are entitled to the same concern for their religious beliefs (true or false, reasonable or unreasonable) as all others. We cannot ignore their concerns because we think of them as “just Sikhs.”

A third problem is that state acts need justifying to all citizens, not just to those whose actions are directly affected by them. One kind of disadvantage in a democracy is not having what you take to be the best policy enacted into law. On many topics, there are opposed views on what the best policy is. Abortions cannot be both permitted and banned: but some people think that, all things considered, they should be banned and others think that, at the end of the day, they should be permitted. Nagel’s view is that a pro-life policy is not neutral because under it pro-choice women will sometimes be coerced into not having abortions; and we cannot offer them reasons they should accept for our coercion. (Pro-choice policies leave the decision up to the mother: so no one is coerced and so the state need not offer anyone reasons.) The thought, presumably, is that justifying pro-life legislation involves an appeal to religious grounds that are controversial, and so reasonable pro-choice women can deny that they have been offered neutral rea- sons. But many pro-life people will rightly feel that there is something fishy here. It is true that a pro-choice regime coerces no one into having an abortion; but it is also true that a pro-choice regime is possible only if we put aside the admittedly controversial considerations that are offered by those who are pro-life. Nagel thinks that questions about the justification of state action arise with especial intensity when peo- ple are being coerced. This seems to me right. But, as I said, laws need justifying to all citizens, not only to those who are coerced by them. And the pro-life person is likely to feel that ignoring her appeal to the sanctity of the life of the unborn fetus is not a way of being neutral between her religious views and others, but a way of taking sides

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

96 u C H A P T E R T H R E E

against them. She can reasonably ask, “Would my views have been ruled out as a basis of state action if they had not been the views of an evangelical Christian (or a Catholic)?” And unless the answer is no, the policy will not count as neutral by the standard of equal re- spect, even though it satisfies Nagel. (I think the answer is no, by the way; so I don’t think our current regime does in fact violate neutrality as equal respect.)

But there are perplexities here we’d better not skate over. Under neu- trality as equal respect, the right question to ask in deciding whether a state act is neutral, in the way that matters, is, Would this person have been treated better had he or she not been regarded as an L? In some cases, it’s easy to make sense of this question. If the San Antonio Spurs are playing the Chicago Bulls and suspect the referee is biased against their team, it would be natural for a Spurs point guard to ask himself, Would that foul have been called if I had been playing for the Bulls? The legal scholar David A. Strauss, with cases of racial and sexual dis- crimination in mind, recommends the test of “reversing the groups.”52

In many cases, though, it’s difficult to know what to reverse—to know what counterfactual we ought to be considering, let alone how we should evaluate it. Consider Thomas v. Review Board, in which the U.S. Supreme Court considered the case of Mr. Thomas, a Jehovah’s Witness who quit his job after being transferred by his employer from a foundry, where steel was prepared for a variety of uses, into a job manufacturing turrets for military tanks. (I owe this example to my colleague Chris Eisgruber.) He decided, after some struggle with his conscience, that his religion required him to be a pacifist and that his new job offended one of his deeply held religious convictions. So he quit. He then applied for unemployment benefits from the state of Indiana and was denied. The Supreme Court, however, invoking Sherbert, sustained his claim for compensation. Writing for the majority, Chief Justice Burger said: “The mere fact that the petitioner’s religious practice is burdened by a governmental program does not mean that an exemption accommo- dating his practice must be granted. The state may justify an inroad on religious liberty by showing that it is the least restrictive means of achieving some compelling state interest.” Burger then rejected the state of Indiana’s claims that there was such an interest and went on to argue that, provided that granting Thomas his unemployment pay did not

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 97

amount to establishment (which, he briskly, and correctly, maintained, it did not), he should therefore have been paid.53

Does neutrality as equal respect offer guidance here? As Chris Eis- gruber points out, had Thomas been a secular pacifist, he isn’t likely to have fared better—on the contrary. On the other hand, Eisgruber observes, “Thomas might have been treated better if he had been a mainstream religious worshiper, forced to quit his job because of a mainstream Christian conviction (e.g., don’t work on Sunday)—or if he had to quit his job because he was transferred to some division (say, textiles) that triggered serious allergic reactions.”

There are two important points raised here. One is a logical point. Evaluating a counterfactual about what would have happened if P is different from evaluating a counterfactual about what would have hap- pened if both P and Q. If it had rained yesterday, I would have stayed home. But not if my mother had needed visiting in the hospital. The question about Thomas should, I think, be whether he would have been granted unemployment pay if he had quit his job for a reason of conscience other than a Jehovah’s Witness’s. The only circumstance we are required to alter in the counterfactual situation is the religious iden- tity under which the relevant state actors conceived of him.

But this raises the second real difficulty, which is that it can be hard in many cases to answer this question. As Eisgruber’s question (and Strauss’s test) suggest, we naturally explore this issue by asking how he would have been treated not if he had been just not-a-Jehovah’s-Witness, but if he had been something else specifically: an atheist, a member of a mainstream denomination, a person allergic to something in the tank division. The answer will then be sometimes better (for a mainstream denomination), sometimes worse (for a conscientious atheist). This does not mean it is indeterminate whether he would have fared better had he not been a Jehovah’s Witness.54 But it does mean that we have to be careful to ask the right question; it is also true that, once we have asked the right question, we may still not be able to answer it clearly.

On the face of the necessarily caricatural account of the facts of the Thomas case given in the Supreme Court’s decision, I think it is very hard to say whether or not the initial finding of the Indiana review board was consistent with neutrality as equal respect. There does not seem, on the face of the record, to be any evidence that Thomas’s reli-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

98 u C H A P T E R T H R E E

gious scruples were treated with anything other than respect because they were seen as the scruples of a Jehovah’s Witness. When the Indiana Court of Appeals reversed, they did so, in essence, because they thought of his case as being like the Sabbatarian cases: and when the Supreme Court of Indiana reversed them (by 3 to 2) the majority seems to have thought that the fact that other Jehovah’s Witnesses regarded work on armaments as permissible was grounds for seeing his decision to quit as “personal” rather than religious. Though these various bodies reached different conclusions, Thomas’s faith as a Jehovah’s Witness does not seem to be why he lost before the two tribunals that rejected his claim.

From the fact that the Indiana board was behaving consistently with neutrality as equal respect, it does not follow, of course, that we should defend the outcome. Their job was to interpret a law in a nonprejudicial way. But we can also ask whether the law itself was well conceived and, in particular, whether the law itself displayed neutrality as equal respect. If you are going to give people unemployment benefits when they lose their jobs or have to give them up for a serious reason—such as a threat to their health—should you also do so when that serious reason is, like Thomas’s, a reason of conscience? To this question I am inclined to say, Yes, absent a compelling state interest to the contrary. But is this required by neutrality as equal respect? Here I think the answer is that what is required is only that you treat equally conscientious reasons equally. More precisely, if you grant an exemption to a member of one denomi- nation who has a reason of conscience substantial enough to make her leave a job, then you should grant it to those whose consciences are driven by other identities. Of course, it can be hard to compare burdens across religions: is a Muslim or Jew who must eat pork in prison more burdened than a secular vegetarian convinced of the rights of animals? There are, I think, no easy answers here. If, however, a state denies (as it might) that leaving a job for a reason of conscience should entitle you to unemployment compensation, I don’t think it offends against equal respect, or against neutrality among identities properly understood. Whether such a decision is wise is a different question.55

There might still be a good liberal reason for framing and interpret- ing the law so as to grant Mr. Thomas compensation: namely, that to do so is to help him sustain his individuality. To give up your job is often a substantial loss, and finding a new one requires that you be able

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 99

to put bread on the table while you search. Unemployment payments to those who leave for reasons of conscience are a way of removing one obstacle to their acting in conformity with their deepest beliefs; one barrier to their sustaining their individualities. Of course, as the state of Indiana observed, there are moral hazards in allowing people to claim conscience. No doubt lazy people could invent all kinds of “conscien- tious” objections. But the rate of unemployment pay in Indiana is surely not sufficiently attractive for this to be a major problem. And there is, as I say, a problem facing us if we do not allow conscientious objection to working on certain projects: namely, that it can require people to choose between remaining in an occupation that betrays their legitimate sense of who they are and putting their families on the street.

To be sure, you should not suppose that you will readily end up with America’s First Amendment regime without reference to the specific history of state and sects that has made the subject so charged. Some liberals, in their theorizing, still try to subsume religious accommoda- tion under a general concern for personal autonomy, say, and avoid special treatment of claims that issue from religious conviction. (The Sikh who wanted an exemption from a helmet law would simply be treated as someone who really, really wanted to keep his turban on.) It’s just that this way of proceeding will not get you to where we are today. You’d come a little closer if you took special account of religious practices as likely to represent deeply constitutive aspects of people’s identity, rather than something like a taste for one candy over another. We can make distinctions between the Mr. Thomases who, with Luther, declare, Ich kann nicht anders, and the Mr. Bartlebys who simply “prefer not to.” We can, even if we are the stoniest of secularists, make distinc- tions between Communion wafers and Necco wafers.

THE LANGUAGE OF RECOGNITION

Talk of neutrality among identities, in ways we’ve seen, helps flesh out what happens when we try to generalize the First Amendment paradigm to deal with identity groups tout court. A state that “established” a certain identity might be taken to have betrayed the aim of neutrality as equal respect, and so might a state that unneces-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

100 u C H A P T E R T H R E E

sarily burdened members of a certain identity group, thus infringing on “free exercise” or its identitarian analogue.56 Of course, the state is constantly imposing differential burdens on this or that identity: the heir to the duke of Omnium will rightly see estate taxes as constraining his life as a scion of a landed aristocracy. But we might say (in the spirit of equal-respect neutrality) that government should not aim, in taxing him, to constrain his identity as such. As this example suggests, the accommodation of identity quickly runs into the problem of the enor- mous scope of that term “identity”: identities are multiple and overlap- ping and context-sensitive, and some are relatively trivial or transient. That’s why those who want to secure state deference (whether affirma- tive or exemptive) toward nonreligious identities take pains to distin- guish the sort worth attention, and the rigorous tests they’ve had to meet: so it is, as we shall see, with Margalit and Raz’s notion of the “encompassing group,” and with Kymlicka’s “societal culture.”

As it happens, at least one generalization of the accommodationist movement has been argued for, in some detail. It proceeds under the rubric of “recognition.”57 It is among the most influential developments in recent political theory, and so perhaps deserving of extended consid- eration. What Charles Taylor has called “the politics of recognition” has little time for either neutralism or autonomy as such: his value term is “authenticity.” And yet the proposal does occupy the same political space, so to speak, as other efforts at brokering a relationship between identities and the state.

The concept of recognition, as it is found in much recent multicul- turalist discussion, is at root a Hegelian one, drawing on the now too familiar discussion of lordship and bondage in The Phenomenology of Spirit. There the thought is that my identity as master is constituted in part by the acknowledgment of my status by the bondsman (and, of course, if less interestingly, vice versa). I cannot be a master, act as and think of myself as a master, unless the bondsman acts toward me as bondsman to master and treats me as a master. It will be uncontrover- sial among those who have normal human relations that the responses of other people play a crucial role in shaping one’s sense of who one is. As Charles Taylor has put it: “On the intimate level, we can see how much an original identity needs and is vulnerable to the recognition given or withheld by significant others. . . . Love relationships are not

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 101

just important because of the general emphasis in modern culture on the fulfillment of ordinary needs. They are also crucial because they are crucibles of inwardly generated identity.”58

But to grasp all this is not yet to say what role the state should have in the regulation of such acts of Hegelian recognition and misrecogni- tion. On the one side lies the individual oppressor whose expressions of contempt may be part of who he or she is, and whose rights of free expression are presumably grounded, at least in part, in the connection between individuality and self-expression. On the other, the oppressed individual, whose life can go best only if his or her identity is consistent with self-respect. How, if at all, is the state to intervene? Some writers, Charles Taylor among them, seem to hold that the state itself, through government recognition, can sustain identities that face the danger of self-contempt imposed by the social contempt of others.

“Policies aimed at survival actively seek to create members of the community, for instance, in their assuring that future generations con- tinue to identify as French-speakers,” Taylor writes, reflecting on the language policies of Quebec. And he insists that the desire for survival is not simply the desire that the social forms that give meaning to the lives of currently existing individuals should continue for them, but requires the continued existence of a way of life through “indefinite future generations.” Yet, though the desire that an identity shall be maintained is not a negligible one, it has to be conditioned and con- toured by other considerations, including the requirements of partici- pation in a larger polity. A politics of recognition, in short, must be buffered by a recognition of politics.

If we are reluctant to accept, as a rationale for Quebec’s language policy, an overriding value of survival, how else might one negotiate the issue of language politics raised by multilingual societies? Let me sketch a way to think about language and schooling in this context, putting to one side the merits of the actual practice that Taylor was interpreting. (Here as elsewhere, I’m happy to put myself in the position of the prover- bial philosopher who demands, “That’s all very well in practice, but will it work in theory?”) Citizenship, we can agree, is one of the primary means for the making of lives in the modern world. The exercise of citizenship requires the capacity to participate in the public discussion of the polity, and so there needs to be a language that is one of the

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

102 u C H A P T E R T H R E E

instruments of citizenship. We can call this the political language. Some states may elect to try to manage more than one political language, and there are complications, to which I will return, that follow from this. But let me pursue the issue first in terms of a single political language.

Public education should aim to give the political language to all citi- zens—and where private education is permitted, it should be required, for the sake of the child who will become a citizen, that mastery of the political language be one of its elements. Such mastery is, of course, consistent with another language’s being the medium of instruction and with learning yet other languages. Our real concern here is what languages the school’s students end up knowing.

Language minorities have an interest in their children’s mastery of the political language; but typically they also have a concern, rooted in the connection between language and identity, that their children should master their “own” language, too. The availability of the minor- ity language is to a great degree a condition for the exercise of one possible identity option, namely, to live a life in which one’s experience as a member of the group is shaped, interpreted, mediated by its lan- guage. This means that children of the minority group for whom this option is to remain open must acquire their language, which might best be accomplished as part of their schooling: and, provided they are also learning the political language, they can thus retain the option without being trapped in a minority identity they cannot escape. So in a country with language minorities, the state should make such options available to parents and children who seek them, if it can. I say “if it can,” because there are resource issues involved in such provision, and trade-offs to be made among many competing desiderata. If the group is very small or the costs would be very high, we may well leave it to the group to sustain the language. The crucial point is that there is a legitimate iden- tity-derived interest here in maintaining the language; the sort of inter- est that a state that cares for its citizens will certainly bear in mind.

There is then, in my view, nothing wrong in the insistence of a state, like Quebec, that all its schools produce competence in the political language. It is, in fact, the right thing to do. This means, however, that the right of anglophone parents to send their children to English-me- dium schools is reasonable only if those schools also teach the political language, i.e., French; and, equally, in francophone schools generally,

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 103

anglophone children (like speakers of other minority languages) should be able—within the limits set by resource constraints that I mentioned earlier—to learn literate mastery of their own language as well.

The complication, of course, is that Quebec is not an independent sovereign but a province of the Canadian nation. That nation has not one but two political languages. As citizens of the Canadian state (and not just of Quebec) francophones are entitled to access to its political language. The question, then, is whether, in states that are officially multilingual, speakers of one of the languages need fluency in the others to achieve equal citizenship. The history of Switzerland suggests that, as an empirical matter, this is not necessary, and in South Africa, for example, with eleven political languages, it would not even be possible. The answer will depend on many historical contingencies (such as what degree of trust between communities exists in view of their past rela- tions), and on questions of political structure (such as the degree to which there is—as in Canada, Belgium, or Switzerland—a decentral- ized politics of regions with different political languages).

And so, whatever the practical merits or difficulties of the current political resolution of these issue in Quebec, the way that it has been conceptualized, via a politics of recognition, gets things exactly the wrong way round. Taylor says, with approval, that “restrictions have been placed on Quebeckers by their government, in the name of the collective goal of survival.”59 But the right issue in deciding the primary language of instruction in state schools is not the maintenance of a francophone ethnic identity—not survivance—but equality of citizen- ship in a francophone state. Under Quebec law, “le français est la langue officielle du Québec.” Once a democratic process has made French the political language, access to French is the right of all citizens and so must be made available to them; access to minority languages is a cen- tral enough interest that children of minority communities should have access to their minority language as well. (And a wise and cosmopolitan community will also encourage children to learn languages that are not, in this sense, theirs.)

Survivance is aided by the choice of French as the political language, of course; that is part of why it was democratically chosen as the politi- cal language. Such an aim is a perfectly acceptable consideration in democratic politics. But such aims must be managed within the frame-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

104 u C H A P T E R T H R E E

work of equal citizenship and a concern for the personal autonomy of citizens, not by notions of compulsory identities. Allowing for the existence of a minority without proper access to the political language offends against equality; placing obstacles in the way of learning other languages offends against autonomy.

If English were the sole political language of Canada, it would be clear, too, that Quebeckers had a right to be taught it as well. But it isn’t; and, given that the country is bilingual, so long as monolinguals are able to exercise full citizenship—a question that is, by its nature, not one that admits of resolutions with sharp boundaries—there is no obvious reason why Quebeckers whose parents don’t want them to learn English should be forced to do so. Of course, they will have wider opportunities if they learn English; they might have even wider ones, however, if they learned Chinese, Hindi, or Spanish. Were it unneces- sary to learn English for adequate political participation, the fact that it provides opportunity would be no more reason to force it on them than the fact that Chinese does. If Quebec officials would be wrong to allow anglophone children to escape French altogether, they would also be wrong to discourage a child from learning English in order to de- prive him or her of access to an anglophone identity. Where people can gain access to an identity by learning a language and they wish for that access, it is not the state’s business to stop them.

The general point is this: there are two ways to bring full citizenship to minority-language communities. One is to make their language one of the political languages, and entitle them to access to official commu- nications in it. (Here you want to be sure that the practicalities of this process will produce real equality of political participation.) The other way is to teach them the political language, while allowing them, if they choose, to maintain their own, which is the route that has been followed in India and the multilingual countries of Africa. The Canadian case makes clear that where political participation is layered and the differ- ent levels have different political languages, these principles may lead to complex outcomes: given the enormously variegated patterns in which languages are distributed across states, however, any solution is going to have to be complex.

The politics of language is a central and difficult issue in the manage- ment of many modern multilingual states, states whose plurality of

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 105

languages result from histories of migration, both voluntary and invol- untary, and of conquest.60 I hope I’ve indicated a plausible way to frame this issue. But after even this crayon-on-a-napkin discussion of policy it is worth repeating a point I made in the preface: this book aims to pick out and explore certain key concepts in our thinking about the ethics of identity. It is not—it does not aim to be—a book of political prescriptions or policy proposals. If what I have said about Quebec is wrong, I hope it is wrong because I have misunderstood the political history or the facts on the ground. What I am committed to is not the policy I have gestured toward, with all its factual predicates, but rather the claims about the way in which considerations of language as a mate- rial for identity and as a tool of citizenship should be brought to bear in the construction of a policy. Here, as everywhere, my aim is to begin with the interests of individuals and to show how identities give individ- uals complex interests that ethics—and, therefore, a satisfactory poli- tics—must bear in mind.

THE MEDUSA SYNDROME

Ethical individualism, it may be worth spelling out, has no simple friend-or-foe relation to “recognition”; and careful readers of Hegel, such as Charles Taylor and Axel Honneth, are surely right that much of modern social and political life turns on such questions of recognition. In our liberal tradition, of course, we see recognition largely as a matter of acknowledging individuals and their identities: and we have the notion, which comes (as Taylor says) from the ethics of authen- ticity, that, other things being equal, people have the right to be acknowl- edged publicly as what they already really are. It is because someone is already authentically Jewish or gay that we deny him something in re- quiring him to hide this fact, to “pass,” as we say, for something that he is not. As has often been observed, though, the way much discussion of recognition proceeds is at odds with the individualistic thrust of talk of authenticity.61 In particular, attending to the oppositional aspects of authenticity would complicate the picture, because it would bring sharply into focus the difference between two levels of authenticity that the contemporary politics of recognition seems to conflate.

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

106 u C H A P T E R T H R E E

To bring out the problem, let me start with a point Taylor has made about Herder—that Herder “applied his conception of originality at two levels, not only to the individual person among other persons, but also to the culture-bearing people among other peoples. Just like individuals, a Volk should be true to itself, that is, its own culture.”62

After all, in many places nowadays, as I suggested earlier, the individual identity, whose authenticity cries out for recognition, is likely to have what Herder would have seen as a national identity as a component of its collective dimension. It is, among other things, your being, say, an African American that shapes the authentic self that you seek to express. And it is, in part, because you seek to express your self that you seek recognition of an African American identity. This is what makes prob- lems for Lionel Trilling’s notion of the “opposing self”: for recognition as an African American means social acknowledgment of that collective identity, which requires not just recognizing its existence but actually demonstrating respect for it. If, in understanding yourself as African American, you see yourself as resisting white norms, mainstream Amer- ican conventions, the racism (and, perhaps, the materialism or the indi- vidualism) of “white culture,” why should you at the same time seek recognition from these white others? (I will have more to say about such paradoxes of support in the next chapter.)

There is, in other words, at least an irony in the way that an ideal of authenticity—you will recognize it if I call it the Bohemian ideal— requiring us to reject much that is conventional in our society, is turned around and made the basis of a “politics of recognition.”63 Now, you may be skeptical of the Bohemian ideal, or see it as a mere indulgence or affectation; but the notion that identities are founded in antago- nism—recall the Rattlers and the Eagles—should by now be an unsur- prising one.

I used the example of African Americans just now, and it might seem that this complaint cannot be lodged against an American black nation- alism: African American identity, it might be said, is shaped by African American society, culture, and religion. Here is how the argument might be framed: “It is dialogue with these black others that shapes the black self; it is from these black contexts that the concepts through which African-Americans shape themselves are derived. The white soci- ety, the white culture, over against which an African-American nation-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 107

alism of the counter-conventional kind poses itself, is therefore not part of what shapes the collective dimension of the individual identities of black people in the United States.”

This claim seems to me to be simply false. What shows that it is false is the fact that it is in part a recognition of a black identity by “white society” that is demanded by nationalism of this form. And “recogni- tion” here means what Taylor means by it, not mere acknowledgment of one’s existence. African American identity (like all other American ethnoracial identities) is centrally shaped by American society and in- stitutions: it cannot be seen as constructed solely within African Ameri- can communities, any more than whiteness is made only by whites.

There is another error in the standard framing of authenticity as an ideal, and that is the philosophical realism (which is nowadays usually called “essentialism”) that seems inherent in the way questions of au- thenticity are normally posed. Authenticity speaks of the real self buried in there, the self one has to dig out and express. It is only later, after romanticism, that the idea develops that one’s self is something that one creates, makes up, like a work of art. For reasons touched on in chapter 1, neither the picture in which there is just an authentic nugget of selfhood, the core that is distinctively me, waiting to be dug out, nor the notion that I can simply make up any self I choose, should tempt us. As we saw, we make up selves from a tool kit of options made available by our culture and society. We do make choices, but we don’t, individually, determine the options among which we choose. To neglect this fact is to ignore Taylor’s “webs of interlocution,” to fail to recognize the dialogical construction of the self, and thus to commit what Taylor calls the “monological” fallacy.

If you agree with this, you will wonder to what extent we should acknowledge authenticity in our political morality: and that will de- pend, surely, on whether an account of it can be developed that isn’t monological. It would be too large a claim that the identities that cry out for recognition in the multicultural chorus must be monological. But it seems to me that one reasonable ground for suspicion of much contemporary multicultural talk is that the conceptions of collective identity they presuppose are indeed remarkably unsubtle in their un- derstandings of the processes by which identities, both individual and collective, develop. And I am not sure whether Taylor would agree with

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

108 u C H A P T E R T H R E E

me that collective identities disciplined by historical knowledge and philosophical reflection would be radically unlike the identities that now parade before us for recognition, and would raise, as a result, ques- tions different from the ones he addresses. In a rather unphilosophical nutshell: my suspicion is that Taylor is happier with the collective iden- tities that actually inhabit our globe than I am: and that may be one of the reasons why I am more hesitant to make the concessions to them that he does. For an ethics of identity (to anticipate my discussion in chapter 5) must confront two distinct though not wholly separable questions: how existing identities should be treated; and what sort of identities there should be.

As we saw, the large collective identities that call for recognition come with notions of how a proper person of that kind behaves: it is not that there is one way that gay people or blacks should behave, but that there are gay and black modes of behavior. These notions provide loose norms or models, which play a role in shaping the ground projects of those for whom these collective identities are central to their individual identities. Collective identities, again, provide what I have been calling scripts: narratives that people use in shaping their pursuits and in tell- ing their life stories. And that is why, as we’ve seen, the personal dimen- sions of identity work differently from the collective ones.

How does this general idea apply to our current situation in the West? We live in societies in which certain individuals have not been treated with equal dignity because they were, for example, women, homosexu- als, blacks, Catholics. Because, as Taylor observes, our identities are dialogically shaped, people who have these characteristics find them central—often, negatively central—to their identities. Nowadays there is widespread agreement that the insults to their dignity and the limita- tions of their autonomy imposed in the name of these collective identi- ties are seriously wrong. One way the stigmatized have responded has been to uphold these collective identities not as sources of limitation and insult but as a central and valuable part of what they are. Because the ethics of authenticity requires us to express what we centrally are, they move, next, to the demand that they be recognized in social life as women, homosexuals, blacks, Catholics. Because there was no good reason to treat people of these sorts badly, and because society contin- ues to provide degrading images of them nevertheless, they demand

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 109

that we work to resist the stereotypes, to challenge the insults, to lift the restrictions.

These old restrictions suggested life-scripts for the bearers of these identities, but they were, in substantial part, negative ones. I need hardly repeat that one does not construct a social identity ab ovo, that our choices are at once constrained and enabled by existing practices and beliefs; but neither do we always “play it as it lays.” And there have been historical moments where we see groups contesting and trans- forming the meaning of their identities with seismic vigor. Certainly this has been a notable dimension of the grand identity movements of the late twentieth century. In order to construct a life with dignity, it has seemed natural to take the collective identity and construct positive life-scripts instead. An African American after the Black Power move- ment takes the old script of self-hatred, the script in which he or she is a nigger, and works, in community with others, to construct a series of positive black life-scripts. In these life-scripts, being a Negro is recoded as being black: and for some this may entrain, among other things, refusing to assimilate to white norms of speech and behavior. And if one is to be black in a society that is racist, then one has constantly to deal with assaults on one’s dignity. In this context, insisting on the right to live a dignified life will not be enough. It will not even be enough to require that one be treated with equal dignity despite being black: for that would suggest that being black counts to some degree against one’s dignity. And so one will end up asking to be respected as a black.

Let me rewrite this paragraph as a paragraph about gay identity: An American homosexual after Stonewall and gay liberation takes the old script of self-hatred, the script of the closet, and works, in community with others, to construct a series of positive gay life-scripts. In these life-scripts, being a faggot is recoded as being gay: and this requires, among other things, refusing to stay in the closet. And if one is to be out of the closet in a society that deprives homosexuals of equal dignity and respect, then one has constantly to deal with assaults on one’s dig- nity. In this context, the right to live as an “open homosexual” will not be enough. It will not even be enough to be treated with equal dignity despite being homosexual: for that would suggest that being homosex- ual counts to some degree against one’s dignity. And so one will end up asking to be respected as a homosexual.

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

110 u C H A P T E R T H R E E

I hope I seem sympathetic to the stories of gay and black identity I have just told, distilling those identity movements of the 1960s and 1970s. I see how the story goes. It may even be historically, strategically necessary for the story to go this way. But I think we need to go on to the next step, which is to ask whether the identities constructed in this way are ones we can be happy with in the longer run. Demanding respect for people as blacks and as gays can go along with notably rigid strictures as to how one is to be an African American or a person with same-sex desires. In a particularly fraught and emphatic way, there will be proper modes of being black and gay: there will be demands that are made; expectations to be met; battles lines to be drawn. It is at this point that someone who takes autonomy seriously may worry whether we have replaced one kind of tyranny with another. We know that acts of recognition, and the civil apparatus of such recognition, can some- times ossify the identities that are their object. Because here a gaze can turn to stone, we can call this the Medusa Syndrome. The politics of recognition, if pursued with excessive zeal, can seem to require that one’s skin color, one’s sexual body, should be politically acknowledged in ways that make it hard for those who want to treat their skin and their sexual body as personal dimensions of the self. And personal, here, does not mean secret or (per impossible) wholly unscripted or innocent of social meanings; it means, rather, something that is not too tightly scripted, not too resistant to our individual vagaries. Even though my race and my sexuality may be elements of my individuality, someone who demands that I organize my life around these things is not an ally of individuality. Because identities are constituted in part by social conceptions and by treatment-as, in the realm of identity there is no bright line between recognition and imposition.

LIMITS AND PARAMETERS

In a well-known essay, “Equality and the Good Life,” Ronald Dworkin takes up Aristotle’s view that “a good life has the in- herent value of a skillful performance,” and proposes what he calls “the model of challenge.” The model “holds that living a life is itself a perfor-

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 111

mance that demands skill, that it is the most comprehensive and im- portant challenge we face, and that our critical interests consist in the achievements, events, and experiences that mean that we have met the challenge well.” Now, the notion I want to borrow from Dworkin is a useful distinction between different ways in which our circumstances figure in the evaluation of how well we have met the challenge. Some of our circumstances (including our own physical, mental, and social attributes) act as parameters, he says, defining what it is for us to have lived a successful life. They are, so to speak, part of the challenge that we must meet. Others are limits—obstacles that get in the way of our making the ideal life that the parameters help define.

Each person, in thinking about her own life, must decide how to allocate her circumstances between these categories, just as an artist must decide which aspects of the tradition she inherits define what her art is and which are barriers to or instruments for her creativity. Dwor- kin writes: “We have no settled template for that decision, in art or in ethics, and no philosophical model can provide one, for the circum- stances in which each of us lives are enormously complex. . . . Anyone who reflects seriously on the question which of the various lives he might lead is right for him will consciously or unconsciously discrimi- nate among these, treating some as limits and others as parameters.”64

Among the circumstances Dworkin regards as his own parameters is his being American. His American-ness is, he says, “a condition of the good life for” him.65 So, for example, even though he has long taught jurisprudence in England and has no doubt influenced the develop- ment of English legal thinking, there surely is, for him, a special signifi- cance to his contributions to American constitutional jurisprudence, a significance that derives from the fact that America—and not En- gland—is his country. So when we’re describing the parameters of our lives, social identities are one obvious class of candidates.

To refer back to the discussion of identification with which I started, you might think that to identify as an L is to treat one’s being an L as a parameter. But just as with identities there is no bright line between recognition and imposition, the relation between parameters and limits is a fluid and shifting affair. Consider homosexuality once more. For some people, homosexuality is a parameter: they are openly gay, and—

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

112 u C H A P T E R T H R E E

happy or unhappy, rich or poor—the life they seek to make will be a life in which relationships with members of their own sex will be cen- tral. Others think of their sexuality as a limitation: they want desper- ately to be rid of homosexual desires, and, if they cannot be rid of them, they would at least like to succeed in not acting on them. At the same time, circumstances that one might assume would be merely impedi- ments may be transmuted into a positive way of being. Thus for many deaf people deafness is not a limit but a parameter: they are not trying to overcome a disability; they are trying to live successful lives as the hard-of-hearing people that they are. A condition becomes an iden- tity—the deaf become the Deaf.

I think the limit-parameter distinction helps us see why “identity” has become a locus of such warring political intuitions. Black, woman, gay, aboriginal—so many of the identity categories that are politically salient are precisely ones that have functioned as limits, the result of the attitudes and acts of hostile or contemptuous others. Each of these categories has served as an instrument of subordination, as a constraint upon autonomy, as, indeed, a proxy for misfortune. Some identities, we can show, were created as part of a classificatory system for oppres- sion. And in the context of antidiscrimination law, say, these identities are treated as a sort of handicap, to be disregarded or remedied. Yet the reversible-raincoat nature of these terms is demonstrated by the fact that categories designed for subordination can also be used to mobilize and empower people as members of a self-affirmative identity. (The disconcerting ease with which limits become parameters recalls the duck-rabbit oscillation between structure and agency we explored in the previous chapter, and it may help explain the ambivalence com- monly occasioned by talk of “identity politics.”) As a parameter, identi- ties provide a context for choosing, for defining the shape of our lives, but they also provide a basis for community, for positive forms of soli- darity. And it is perfectly consistent to consider your membership in an identity group as both parameter and limit: the Black Nationalist who deplores white supremacy is keenly aware of color as constraint, even as he seeks to make it the basis of a political mode of resistance. There is a two-way traffic between limits and parameters, then.

Yet to say all this is not to lose sight of other aspirations and other ideals, among them what Peter Singer has called the “expanding circle”

This content downloaded from 154.59.124.59 on Mon, 19 Oct 2020 06:21:45 UTC

All use subject to https://about.jstor.org/terms

T H E D E M A N D S O F I D E N T I T Y u 113

of our moral sympathies. The contours of identity are profoundly real: and yet no more imperishable, unchanging, or transcendent than other things that men and women make. Indeed, talk of division and dishar- mony can lead us to neglect the powerful countervailing forces: conflict may be productive of identities, but conflict can also be a powerful uni- fying force across identity groups. It is not a new thought—and in some ways not an entirely heartening one—that the two world wars had a powerful effect in forging an American identity that lessened the social salience of small social divisions; that the Second World War, in particu- lar, did as much as anything to make the civil rights era inevitable.

What happened at Robbers Cave State Park confirms this intuition with almost comic schematism: the intergroup bloodlust that was so easily conjured into being finally subsided before what the researchers called “shared superordinate goals.” As the animosity between the groups and solidarity within them was beginning to boil over, the inge- nious researchers devised a series of communal crises. Necessity was the mother of amity.

To start with, the researchers sabotaged the camp’s water system. Two groups of increasingly thirsty boys worked together to identify the problem and repair it. And then a truck, which was to have fetched food for them, broke down, or gave every appearance of having done so. It would have to be pulled up with a rope; and only the collective force of both groups—the raffish Rattlers and the high-minded Ea- gles—would suffice. By the by, banners were set down, putative differ- ences set aside. Shoulder to shoulder, tugging on their tug-of-war rope (the experimenters had an appreciative eye for symbolism), the kids winched the stalled truck uphill.

No single crisis was enough to erase the poisonous animosity be- tween the Rattlers and the Eagles; a series of them, however, produced genuine confraternity and the dissolution of social boundaries. The truck started up, food was brought back, and a meal was collaboratively prepared. “After supper a good-natured water fight started at the edge of the lake,” the researchers dutifully noted; but, as they were careful to add, “the throwing and splashing was not along group lines.”66

This content downloaded from 154.59.ffff:ffff:ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC

All use subject to https://about.jstor.org/terms